U.S. Data Center Demand Could Surge to 106 GW by 2035 — What It Means for Power Planning and Infrastructure
U.S. power demand from data centers is poised to grow dramatically over the next decade, with new analysis suggesting requirements could reach 106 gigawatts (GW) by 2035 — a level that would make data centers one of the largest categories of electricity use in the country. This projection, reported by BloombergNEF and highlighted in a recent Utility Dive article, underscores how digital growth and artificial intelligence workloads are reshaping the future of grid planning, energy procurement, and infrastructure investment.
The projected figure, derived from BloombergNEF’s dataset and industry modeling, reflects a confluence of trends: cloud expansion, hyperscale compute deployments, AI and machine learning workloads, and enterprise digital transformation initiatives. For context, if data center load reached 106 GW, it would rival electricity consumption levels of entire industrial sectors and would require utility planners to rethink how power is produced, transmitted, and managed across regions.
Drivers of Rapid Load Growth
Several factors contribute to the surge in projected demand:
- AI and High-Performance Computing: Generative AI, large language models, and other advanced applications require sustained, high-density compute power that translates directly into increased electricity use.
- Cloud Services and Digital Platforms: As organizations shift workloads to the cloud, global hyperscalers continue to build data center capacity to support everything from storage to edge computing.
- Enterprise Expansion: Large enterprises are expanding their own compute footprints to support analytics, remote services, and bandwidth-intensive applications.
The result is load growth that far exceeds traditional demand patterns seen in residential or general industrial use — which historically dominate many grid forecasts. In parts of the U.S., particularly in states such as Virginia, Ohio, and Texas, data center planning is already pushing planners to update long-range transmission and generation expansion plans to accommodate clustered high-demand users.
Grid Implications and Strategic Planning Needs
If broadcast predictions come to pass, utility planners and grid operators will face distinct challenges in maintaining reliability while expanding capacity. The sheer scale of 106 GW — the rough equivalent of 100 large power stations — would require not only substantial new generation but also significant transmission and distribution upgrades.
One key concern is interconnection timelines. Historically, data center site builds have been tied to long lead times for transmission service agreements and upgrades. In many cases, developers spend years in interconnection queues, only to find that upgrades lag actual project timelines. As data center load projections balloon, so does the complexity of queue management and grid access planning.
In areas with constrained transmission capacity or limited new build ability, developers and operators may increasingly consider on-site generation, hybrid power architectures, or behind-the-meter resources as part of their resiliency and planning strategies. These approaches can help mitigate bottlenecks while supporting reliability for mission-critical loads.
Policy, Investment, and Regulatory Considerations
The implications of such dramatic load growth extend into regulatory and policy domains. If data centers collectively approach—and potentially exceed—100 GW of demand, state utility commissions, regional grid operators, and federal entities such as FERC will be under increasing pressure to create clearer frameworks for planning, cost allocation, and resource adequacy.
Already, grid operators like PJM Interconnection, ERCOT, and MISO are wrestling with how to balance traditional load growth with the concentrated, high-demand nature of data center power use. Questions about how to allocate transmission upgrade costs, how to integrate energy storage or demand response, and how to ensure resilient service across load types are now central to many power planning forums.
For example, PJM has recently seen capacity auction prices climb significantly — a market response that reflects tightening resource conditions partly driven by projected data center loads. In response, policymakers have proposed various ideas such as “load responsibility” frameworks that tie large load connections to contributions toward grid costs or require flexible capacity commitments from large customers.
What This Means for Generator Data Readers
For professionals in the Generator Data ecosystem — including electrical contractors, service providers, generator dealers, and data center power planners — these projections carry clear, actionable signals:
• Project Planning Must Anticipate Scale:
Operators evaluating sites or planning expansions should factor in long-term capacity needs — and recognize that grid delivery may become more constrained without proactive engagement in interconnection planning.
• On-Site and Hybrid Power Strategies Grow in Relevance:
Because interconnection and grid expansion timelines can lag, integrating on-site generation, combined heat and power (CHP), or hybrid storage-generation systems may help meet load commitments while ensuring reliability.
• Cost Allocation and Contracting Conversations Are Evolving:
As states and utilities explore new frameworks for how costs are shared, buyers must stay informed about rate design and interconnection cost structures that could impact long-term energy budgeting.
• Partner Collaboration Becomes Strategic:
Early engagement with contractors, utilities, and service partners can help navigate complex infrastructure demands, enabling smoother project delivery and more resilient operations.
Balancing Growth With Affordability and Reliability
The data center sector’s anticipated trajectory — with demand potentially reaching 106 GW by 2035 — reflects broader trends in digital economy growth, AI compute intensity, and enterprise cloud adoption. Yet for grid operators and power planners, reconciling this pace with reliability standards and affordability for all customers will require new thinking, collaborative frameworks, and innovative approaches to infrastructure development.
For the Generator Data audience, these developments reinforce that power strategy isn’t an afterthought — it’s a central component of successful data center planning and a catalyst for new infrastructure solutions across the industry.
Source
Utility Dive — “U.S. data center electricity demand could reach 106 GW by 2035, BloombergNEF says”
🔗 https://www.utilitydive.com/news/us-data-center-power-demand-could-reach-106-gw-by-2035-bloombergnef/806972/